Showing papers in "Ecological Informatics in 2019"

TL;DR: The maximum-entropy algorithm was applied to predict the current and future potential distribution of Rosa arabica Crep to provide a basis for its protection and conservation and can be used to define the high priority areas for reintroduction or for protection against the expected climate change impacts and future modifications.

TL;DR: This work studies both the fine tuning of several deep learned models and transfer learning from the same models with the aim of exploiting their diversity in designing an ensemble of classifiers, and shows how to combine different CNN in order to improve the performance.

TL;DR: This study reviews the general principles and applications of state space models, hidden Markov models, random forests, and support vector machines in the inference of animal behavior from movement data to improve understanding of environmental conditions on animal movements and behavioral decisions.

TL;DR: Sentinel-2 satellite imagery combined with field-measured biomass using Random Forest (RF), a machine learning regression algorithm, is applied to estimate forest aboveground biomass (AGB) in Yok Don National Park, Vietnam to demonstrate the potential to effectively predict the spatial distribution of forest AGB with adequate accuracy.

TL;DR: Temperature was the most important variable affecting the potential distributions of J. blanfordi and J. loftusi in Iran and predicted the impact of climate change on their future potential distributions using two different modelling software packages: Maxent and sdm.

TL;DR: Seven heuristic methods of variable selection are compared against a novel approach that proposes to select best sets of variables by evaluating performance of models created with all combinations of variables and distinct parameter settings of the algorithm in concert.

TL;DR: This paper investigates acoustic features, visual features, and deep learning for bird sound classification, which results indicate that the proposed deep learning method can achieve the best F1-score 94.36%, which is higher than using the acoustic features approach and using the visual features approach.

TL;DR: The results showed that the SVH is season and sensor dependent, the goodness of the Rao's Q index, the relevance of the NDVI in the study of the SVh and the importance of the multi-temporal approach.

TL;DR: Akaike information criterion should not be used if users are interested in prediction more than explanation in ecological niche modelling and those models with the lowest AIC values tend to generate geographical predictions with high commission and omission errors.

TL;DR: The results showed that RF can be effectively applied to predict the spatial distribution of LAI and k, and followed an inverse relation in accordance with the Beer Lambert's Law.

TL;DR: Analysis is presented that shows how object localization accuracy is increased by an automatic correction mechanism in the deep network's cascaded ensemble structure that rectifies any errors in the predictions as information progresses through the network cascade.

TL;DR: The results of this study can allow land managers to avoid wasted human effort and materials as well as the exhaustion of wild P. ostii resources that could result from the blind introduction of this species into unsuitable habitat while improving both the quality and yield of P. Ostii.

TL;DR: A new CNN composed of three branches that classify the fish species, family and order is proposed with the aim of improving the recognition of species with similar characteristics and results showed that the proposed method provides superior results to traditional approaches.

TL;DR: Ten principles for the current best practice in EBV-focused biodiversity informatics are encapsulated as ‘The Bari Manifesto’, serving as implementation guidelines for data and research infrastructure providers to support the emerging EBV operational framework based on trans-national and cross-infrastructure scientific workflows.

TL;DR: An algorithm based on Gaussian Mixture Models together with Pixel-Wise Posteriors for fish detection in complex background scenarios is presented and yields an F-score of 84.3%, which is the highest score reported so far on the aforementioned dataset for detecting fish in an unconstrained environment.

TL;DR: A framework for analyzing the spatial characteristics of land uses and calculating ecological compensation from 2000 to 2015 in Sichuan Province, China provides a better understanding of spatial characteristic scales of land use and enables evaluation of the ecological integrity of landscapes.

TL;DR: An innovative computer vision application for the identification of unknown Risso's dolphin individuals that relies on a feature-based automated approach relying on SIFT and SURF feature detectors is shown.

TL;DR: This paper demonstrates the application of a convolutional neural network for the automatic detection and classification of woody regrowth vegetation in repeat landscape photographs and tests if the classification results based on the automatic approach can be used for quantifying changes in woody vegetation cover between image pairs.

TL;DR: Combining automatic classification with manual ID through fully customizable classifiers implemented in open-source software as demonstrated here shows great potential to help overcome the acknowledged risks and biases associated with the sole reliance on automatic classification.

TL;DR: This study proposes guidelines on how machine learning can be used for specific applied and theoretical applications in a SDM context and recommends the use of numerical predictions for species distribution modeling since they help to convey more information than binary predictions.

TL;DR: A framework using airborne imagery, object-based image approach (OBIA), hyper-spectral analysis and Random Forest to classify vegetation along narrow, semi-arid riparian corridors via a case study of the Grand Canyon, the Colorado River is developed.

TL;DR: The results showed that the spatial distribution of RE is characterized by it increasing from west to east and north to south, while in most areas, the annual RE has increased slightly, while this upward trend in the southeastern part of the study area was more significant.

TL;DR: The aim of DEIMS-SDR is to be a globally comprehensive site catalogue describing a wide range of sites, providing a wealth of information, including each site's location, ecosystems, facilities, measured parameters and research themes and enabling that standardised information to be openly available.

TL;DR: It was contended that a combination and integration of fractal theory, GIS and RS can substantially advance the study of landscape ecology for data acquisition, process modeling, scale transformation, result analysis and visualization.

TL;DR: Compared with other forest management methods, the SBFM can achieve the goal of optimizing the spatial structure of forests, which is conducive to the growth of trees and the improvement of forest stand productivity.

TL;DR: It is argued that one of the priorities of the local research agenda must be to consolidate the ensemble character of the modelling work in Lake Erie, where the performance of the aquatic ecological models in the Lake Erie declined from physical, chemical to biological variables.

TL;DR: It is argued that selected methods to derive terrain attributes from DTMs should be clearly described to encourage reproducibility and proper interpretation of results, thus enabling a better understanding of the role of scale in ecology.

TL;DR: A new approach for automating building construction when improving their energy efficiency is proposed, aiming to foresee comfort levels based on Heating, Ventilating, Air Conditioning (HVAC), constructive systems performance, environmental conditions, and occupant behavior.

TL;DR: A sliding window algorithm of short frame length is suitable for differentiate the Mel-spectrogram of bird sound and the GRU network is connected and used as a classifier to directly output the prediction results.

TL;DR: Habitat-Net is a novel deep learning application based on Convolutional Neural Networks to segment habitat images of tropical rainforests and reduces the degree of uncertainty that would be introduced by manual processing of images by different people (either over time or between study sites).